Nation: Disco Beat in 1978 Politics

• Share

Polls misread a small and volatile electorate

With all the votes counted and in some cases recounted, politicians, pollsters and pundits—the subculture that makes its living from the expensive game of competing for public office—sat back last week to try to understand not just what had happened but exactly why. What emerged from this post-mortem on the 1978 elections was a vision of American politics in which large numbers of people sat by like wallflowers in a dance hall while small groups of voters performed an erratic sort of disco dance.

The volatile electorate fooled pollsters in a number of states. Pat Caddell, who handled Jimmy Carter's polling in 1976, assured New Hampshire Democratic Senator Thomas Mclntyre that he was leading Gordon Humphrey by 59.5% to 30%, with no signs of movement toward the Republican. Humphrey won, 51% to 49%. Respected Pollster Peter Hart found that incumbent Democrat Dick Clark was leading his conservative Republican opponent Roger Jepsen 57% to 27% in October. "We did not have it tight, and we did not have Jepsen moving up," says Hart. Jepsen beat Clark, 52% to 48%. In Kansas, one survey had Democrat Bill Roy ahead of eventual Winner Nancy Landon Kassebaum; running for Governor of that state, Democrat John Carlin was behind in the polls but also emerged a winner.

More than usual, the analysts were dealing with a shifting and uncertain electorate. "Nothing stabilized," says Hart. "There were no overriding issues. People knew more what they didn't want than what they wanted. They didn't put the total record of an individual into perspective." Observes Political Consultant John Deardourff: "When you add a low threshold of interest to a lack of commitment to candidates or parties, the pollsters have a terrible time."

Right up until Election Day, an unprecedented proportion of voters remained undecided—and then made up their minds on the basis of personality or on one inflammatory issue like abortion. "Large numbers of voters are saying that it doesn't matter which party the candidate belongs to," observes Caddell, "and he can't solve the problems anyway, so personality is predominant. You can see that in the great number of negative 5 campaigns directed by one candidate 2 toward another, and that's the toughest thing to follow."

Part of the pollsters' problem was low turnout, which made their samples less reliable than usual. "People don't like to concede that they don't vote and have no interest," says Deardourff.

"They are much more likely to fool an interviewer on the telephone than during a personal session." Pre-election polls usually rely on one in-depth, in-person survey conducted a few months before the campaign. Follow-up polls are usually done by telephone. Most candidates prefer to save the $50,000 that a person-to-person follow-up would cost. This year that was a critical mistake in many states.